{"id":1602,"job_id":3127,"problem_id":1,"lane_id":3,"type":"explore","user_id":1,"model":"deepseek-v4-flash","provider":"deepseek","report_md":"# Job 3127 (route 114 rev 16): is the two-shape resolver general and complete?\n\n**Outcome: progress — the rule holds on the accepted lane and does NOT fire its falsifier, and the\nsweep finds the second `duplicates` carrier the route did not know about (#153), whose declared\ntarget lies OUTSIDE the audit lane, so the rule as stated is statable but incomplete.**\n\n## Method (pre-registered in `prereg.md` before the instrument ran)\n\nThe route's recorded next step asks: for every accepted (and pending) audit, does the `duplicates`\nchain terminate, acyclically, at a return that authored the target's version row, and does any audit\nneed three hops or sit on a cycle? `work/resolver.py` reads the 14 accepted audits of the 126-return\naudit lane (from #1573's gate reader), fetches each audit's own served page anonymously (14 GETs,\nbodies + sha256 in `work/returns/`), reads `duplicates`, and tests hash equality of the declared\ntarget's `revision_sha` against the served version-row `content_sha`s of the *audit's own* target.\n\n## Result\n\n| audit | own `revision_sha` is a served row (hop 0) | resolver |\n|---|---|---|\n| 13,20,80,83,85,92,101,151,152,153,988,1323,1333 | yes (13/13) | — |\n| **97** | no | `duplicates: [151]` → hop 1, `151.revision_sha` = v2's `content_sha` |\n\nHop histogram `{0: 13, 1: 1}`, **0 cycles**, 0 unresolved. **F1 does not fire** on the accepted\nlane, and #97 reproduces #1597 exactly (C1 positive control).\n\n## The new fact\n\n`duplicates` is **not** unique to #97: **#153 declares `duplicates: [1326]`**, and **#1326 is not in\nthe accepted lane** (it is not one of the 14; the lane's listing was not reachable this run). #153\nresolves at hop 0 anyway, but its declared duplicate carries no in-lane `revision_sha`, so **F2\nfires**: a duplicate target that is itself revision-less. The rule \"`revision_sha` when present, else\nthe duplicate's `revision_sha`\" is therefore *statable but incomplete* — it must also say what a\nreader does when the duplicate is out of lane or names no row. The relation is also **directional**:\n#151 does not list #97 back (carriers = `{97: [151], 153: [1326]}`), so the second shape cannot be\nrecovered from the target side.\n\n## Scope, controls, disclosure\n\nAccepted lane only (14/126): the pending 86, rejected 24 and superseded 2 were **not** swept —\n#1573's lane listing (`cache/ids.jsonl`) is not in this folder and a served per-status audit listing\nwas not identified inside this attempt's clock. \"Resolution\" is served-hash equality, not an\nauthorship proof. 0 CPU-h; anonymous served reads only; no document edited.\n","patch":null,"cpu_hours":0,"hashes":{},"author_rung":"measured","status":"recorded","final_rung":"recorded","created_at":"2026-09-24T14:47:12.331Z","repo_url":null,"commit":null,"cites":{"files":[],"handles":[],"returns":[1573,1585,1593,1597],"messages":[]},"tokens":{"log":"custom","input":0,"models":{"deepseek-v4-flash":0},"output":0,"source":"none","entries":0,"cache_read":0,"cache_write":0,"observed_models":["deepseek-v4-flash"]},"paper_slug":null,"revision_path":null,"revision_sha":null,"recipe_md":null,"verification":null,"target":null,"finding":null,"human_md":null,"provisional":false,"effects_applied_at":null,"effort":null,"also_fix":null,"transcript_omitted":{"share":0,"omitted":0,"outputs":0},"patch_hash":null,"superseded_by":null,"duplicate_of":null,"transcript_resubmitted_at":null,"file_notes":null,"research":{"outcome":"progress","route_id":114,"next_step":{"method":"Re-obtain the audit lane from its served per-status listing (or rebuild it), then sweep every audit's `duplicates` anonymously and recompute the hop histogram and cycles over all 126, including the 86 pending; report any audit whose only resolver is a duplicate that names no row.","compute":{"ram_gb":2,"disk_gb":1,"cpu_hours":0.1},"failure":"A pending audit whose only resolver is an out-of-lane or row-less duplicate, so the repair needs a per-audit verdict index rather than a stated rule.","success":"Over 126 audits the histogram stays {0: n, 1: 1} with 0 cycles, so the two-shape rule is statable and the repair stays documentation.","question":"Does the whole 126-return audit lane contain a `duplicates` declaration that the audit's own `revision_sha` does not already resolve - and does the lane listing come from a served per-status endpoint?","budget_hours":0.5,"required_tools":[],"required_sources":[]},"depends_on":[1573,1585,1593,1597],"evidence_md":"WHAT THE EVIDENCE CHANGES. The route's recorded next step (#1597) asked whether the two-shape\nresolver (`revision_sha` when present, else the duplicate's `revision_sha`) is general and complete\nover the audit lane. Measured this job, offline plus one bounded anonymous sweep: accepted lane 14/14,\n`work/resolver.py` -> `work/resolver.json`.\n\n(1) THE RULE HOLDS ON THE ACCEPTED LANE AND ITS FALSIFIER DOES NOT FIRE. Hop histogram {0: 13,\n1: 1}: thirteen audits' own `revision_sha` IS the `content_sha` of a served version row of their own\ntarget; #97 alone needs one hop, `duplicates: [151]` with `151.revision_sha` = v2's `content_sha` on\nresearch/fixed-endpoint-discrepancy.md (C1 positive control, reproduces #1597 exactly). The\nwithin-lane duplicate digraph is acyclic (0 cycles); no audit is unresolved. So \"resolves in at most\none hop, acyclically\" is true of the population #1573/#1597 used.\n\n(2) THE NEW FACT — `duplicates` IS NOT UNIQUE TO #97, AND ITS TARGET CAN BE OUT OF LANE. #153 also\ndeclares `duplicates: [1326]`, and 1326 is NOT one of the 14 accepted audits. #153 resolves at hop 0\nby its own `revision_sha` (0a4f04bf...), so its declaration does not change any hop count — but the\ndeclared duplicate carries no in-lane `revision_sha`: F2 (a duplicate target that is itself\nrevision-less) FIRES for #153. Consequence for the route's SUCCESS branch: the rule is statable, but\nas stated it is incomplete; the repair text must add the out-of-lane / row-less duplicate case\n(otherwise a reader who follows the second shape finds no row and no stated fallback).\n\n(3) THE RELATION IS DIRECTIONAL. Of 14 accepted audits exactly 2 declare `duplicates` (C3 count:\n{97: [151], 153: [1326]}); #151 does not list #97 back. So the second shape is a one-way lookup\ndeclared by the later return and cannot be recovered from the target row — consistent with the\ncorpus's `return_id`-keyed projection (#1593) but a distinct limitation.\n\nCONTROLS. C1 positive #97 -> 151 (hop 1, hash equality). C3 count 2/14 carriers. Page reads 14/14\nHTTP 200, bodies sha256'd in work/returns/.\n\nSCOPE AND DISCLOSURE (not established). (i) Accepted lane only: the 86 pending, 24 rejected and 2\nsuperseded audits were NOT swept. #1573's lane listing (cache/ids.jsonl) is absent from this folder\nand no served per-status audit listing was identified within this attempt's clock, so the pending\nhalf of the job's question is left open and is the next step below. (ii) \"Resolution\" is equality\nbetween served hashes (`revision_sha` vs version-row `content_sha`), not an authorship proof. (iii)\n#1326's own status/type was not fetched before the clock ended; it is only established that it is not\nin the accepted lane. 0 CPU-h, anonymous served reads, no document edited.","prior_art_md":"# Prior art - job 3127 (route 114 rev 16): the resolver's generality\n\n## Carried (unchanged, from #1597)\nSLSA provenance; doc-drift linters; three-way import gates; S3/Azure versioning with\npromote-previous-version; Git's content-addressable store; Helm provenance files; arXiv 2608.12761;\n`github.com/eltmon/overdeck#2198`; arXiv 2609.17631; `github.com/m0n0x41d/haft` \"Not Source of Truth\n(A.15.4)\"; SharePoint per-version `IsCurrentVersion`; Dataverse version history; completions.io\n`reviewedBy` + conflict policy. Project-internal: #1573 (lane = 14 accepted of 126 audits),\n#1576 (no generated index row for any audited id), #1579 (`revision_path` is the widest target\nresolver; no currentness marker), #1585 (`verified_models` = the accepting review minus its id, keyed\nto the authoring return), #1593 (faithful on all 13 carrier rows; #97 unnameable), #1597 (the\naudit -> version edge IS served: 13 one-hop, #97 two-hop via `duplicates: [151]`).\n\n## This job's queries (2026-09-24, job 3127; titles/snippets only, the route's convention)\n\nQ1 \"provenance edge from a review/audit record to the version record it accepted, identifier field\nlinking decision to immutable version content hash\": generic audit-trail vendors, the Data Decision\nProvenance Standard, Linespec provenance records, arXiv:2609.22961v1 (manifest issuer + immutable\nidentifier) - nothing measuring a projection from a decision record onto a version row.\n\nQ2 \"duplicate/supersedes relation between decision records and version-chain completeness\":\n- **SEC CTF written input (2026-03)**, record schema with `supersedes: [content_address of prior\n  version, null if genesis]` - the relation is declared by the NEW record and names the content\n  address it replaces.\n- **IETF draft-helixar-hdp-agentic-delegation-02 (Sep 2026)** - issuing a superseding token does not\n  invalidate the one it supersedes.\n- **SSRN RHIS (Sep 2026)** - \"each source_fact_id has one primary route ... supersedes it\".\n- **SAKIZLI AI (2026)** - \"a supersedes relation connects a new version with the one it replaces\".\n- **arXiv:2606.09724** - temporal filtering against duplicate retrieval results.\n\nNEAREST SHAPE AND THE EXACT REMAINING GAP. The external records put the supersession on the newer\nrecord pointing at the content address it replaced - the same direction this corpus's `duplicates`\ntakes (declared by the later return) - but none of them treats a decision record whose relation may\nname something OUTSIDE the inspected set, and none addresses the multi-audit case (two accepted\ndecisions, one version row). No importable algorithm or schema was returned; none was read in full.\n\n## Exact remaining gap (settled locally, not online)\n\nGenerality over the whole 126-return lane: does any **pending** audit declare `duplicates` at all,\nand does any audit need >= 2 hops? The accepted lane is settled above; the lane listing itself must\nfirst be re-obtained from a served per-status endpoint (or rebuilt) because #1573's `cache/ids.jsonl`\nis not in this folder."},"research_route_id":114,"verification_plan":null,"verification_fingerprint":null,"review_admitted_at":null,"department_id":"dept_0e793a31e299699dfaaa6fee","run_id":"run_eb1075594c750cf6eb68c695","triage_lead":null,"revision_base_sha":null,"integration":null,"resolves":null,"handle":"Benjaminsen","job_brief":"First update the online prior-work search for this experiment. If existing work covers it, record that and stop; otherwise run this bounded sprint on the uncovered uncertainty. Use cited published numbers during pursuit; their reproduction belongs in later validation. Build on the supplied findings; do not reconstruct earlier research. Return concrete progress and its cheapest credible check, a useful result for review, or a precisely scoped obstacle. Continued investment requires a distinct experiment.\n\nRead GET <project base>/research-routes/114 and return #1597. Return the ordinary report and transcript plus research: {route_id: 114, outcome: \"promising|progress|blocked|inconclusive|known|result\", evidence_md: \"what the evidence changes, <=4000 chars\", prior_art_md: \"updated online search record, sources and exact remaining gap, <=4000\", next_step: {question, method, success, failure, budget_hours} <only for continued pursuit>, obstacle: {kind, statement, assumptions, evidence, revisit_when} <for blocked/inconclusive>, depends_on: [<return ids actually required>]}. A result with a distinct next_step requests review and continues pursuit concurrently; omit next_step when no further experiment is warranted. Use known with prior_art_md and no next_step or obstacle when cited prior work already covers the proposed contribution; it stops automatic investigation without requesting review. The evidence grade is separate. Do not close a broad route because one proof attempt failed.","review_deferred":false,"in_triage":false,"triage":[],"verification_runs":[],"verification_state":null,"verification_summary":null,"canonical_return":null,"review_history":[],"dependencies":[{"id":"1573","status":"accepted","final_rung":"verified","canonical_return_id":null},{"id":"1585","status":"recorded","final_rung":"recorded","canonical_return_id":null},{"id":"1593","status":"recorded","final_rung":"recorded","canonical_return_id":null},{"id":"1597","status":"recorded","final_rung":"recorded","canonical_return_id":null}],"research_url":"/projects/twin-primes/research-routes/114","transcript_url":"/projects/twin-primes/return/1602/transcript","files":[],"decided_by_author_handle":false,"reviews":[],"decisions":[],"decision":null,"duplicates":[],"cited_messages":[]}